Learn about known issues and apply workarounds for adapters supported by Event Stream Processor.
CR# | Description |
---|---|
650777 |
For the Database Input adapter, when data is converted between the database and Event Stream Processor datatypes (numeric) that have different ranges, the JDBC and ODBC drivers may handle these operations differently. For example, the conversion may get rejected or convert into an incorrect value, and so on. |
680661 |
The Platform Input and Platform Output adapters are not supported by Event Stream Processor. The platform_in.cnxml and platform_out.cnxml files are included in the install, but are for internal use only. |
684802 |
To use schema discovery for the Reuters Marketfeed and OMM Input adapters:
|
684922 and 684945 |
When running the Reuters OMM adapter in interactive publish mode, you may experience intermittent stability issues when connected directly to rmdstestclient (a Reuters utility bundled with their server). |
695586 | When using the Database Input adapter
and the getData() built-in SPLASH function in the flex stream, there is
a compatibility issue between JDK7 and the JDBC driver for MSSQL
(3.0.1301.101 and 4.0 CTP3). The problem occurs when mapping a datetime2 datatype into Event Stream
Processor timestamp or string datatypes as the adapter executes
queries such as a SELECT statement. The value that is read in is
incorrect. To work around this issue, use the SQL convert() built-in
function to cast a datetime2 column
into a varchar. For example, modify a
SELECT statement like
this:select c_bigdatetime from newalltypesinto this: select convert(varchar(30), c_bigdatetime, 121) from newalltypeswhere the third argument value of 121 is the format to which the datetime2 field will be converted. |
714685 |
You must make a few changes so that Java external
adapters from Event Stream Processor 5.0 are compatible with ones
from 5.1.
|
715027 | A custom internal adapter that was
written using the 5.1 version of the GenericAdapterInterface.h file is not compatible with an
adapter that was written using the 5.0 version. To make these adapters
compatible in 5.0, change the following function signatures for these
files to the 5.1 version:
|
715811 | When the peer program publishing to
the adapter does not close its socket connection cleanly the adapter's
state becomes DEAD and it no longer accepts connections. To avoid this issue, ensure that the publisher issues a "close" on the socket connection before exiting. |
716339 | The CSV Input adapter does not process the last line of the CSV input file. To ensure that all data in the file is processed, add an empty line at the end of the file. |
716843 | If you stop the Database Input
adapter while it is reading data and then start it again, the adapter
does not start and therefore, does not complete reading data. To work around this issue, restart the ESP Server, and if you did not configure the adapters to start automatically, then manually start the Database Input adapter. |
716844 | If you drop the table that the
Database Input adapter is polling data against, the adapter crashes. To work around this issue, restart the ESP Server, and if you did not configure the adapters to start automatically, then manually start the Database Input adapter. |
717862 | To use an adapter with ESP, you must set the JAVA_HOME environment variable to the location where you have installed Java 1.7 in the SYBASE.sh file. |
721462 | If ESP bigdatetime is mapped to timestamp in HANA,
deletes or updates are processed in ESP, but not in HANA. This is due to
an issue in HANA where the timestamp data type truncates microseconds to
milliseconds. Because the original, non-truncated value is not found in
the database, the delete or update fails. To work around this, if you insert data into HANA through the HANA adapter, issue deletes or updates outside of the adapter. When constructing your delete or update statements, use millisecond precision. |